Friday, September 1, 2017


Talk to VCTA Teachers Day, Melbourne, Friday, September 1, 2017

I want to talk to you about the changing economics of wages and the labour market. Some of what I say may be news to you, some of it will, I hope, bring back to you stuff you haven’t thought about for a long time. What I say is intended more for your own edification – or re-edification – than for you to take straight into the classroom and lay on your students. One of the purposes of Professional Development is, after all, to ensure that you know a lot more about the subjects you teach than your students do.

The neoclassical model

One of the things I’ve noticed in my career as an economic journalist is the gap in thinking between economists who specialise in the study of a particular market or industry – the labour market, for instance, or the health industry, or even the education industry – and other economists who specialise in a different aspect – monetary economics, fiscal policy – or have no particular specialty. The specialists specialise in knowing about all the peculiarities of their market – all the special cases of market failure - that make it different from other markets and much harder to analyse. By contrast, the non-specialists “specialise” in using the same generalised, simple neoclassical model to analyse all markets on the assumption that all markets, being markets where prices change to equilibrate supply and demand, are pretty much the same.

I’ve noticed this one-model-fits-all approach particularly among policy advisers – Treasury, the Productivity Commission, PM&C, the Reserve Bank – but it has become more common since the rise to intellectual dominance in the early 1980s of what we used to call “economic rationalism” and now have joined the rest of the world in calling “neoliberalism”. This move to a more fundamentalist approach to economic analysis was very much about playing down the incidence of market failure and using the same simple, price-driven model to explain everything. Its great attractions are its simplicity, its clear predictions and its clear prescriptions on how problems should be solved. One sign of reversion to a more fundamentalist approach has been the willingness of economists to advocate – or, at the very least, accept in silence – the push for a return to individual contracting between workers and their bosses. Of course, this “neoliberal consensus” is now breaking up before our eyes under the onslaught from Brexit, Trump and the Redheaded One, and I’ll present you with some evidence of changing attitudes in academia and among Australian policy advisers.

The high school economics syllabus contains a far bit about the changing institutional arrangements for wage fixing in Australia, but doesn’t dwell on the micro theory of how wages are set by a firm or an industry. The syllabus’s Keynesian approach to macro management implies acceptance, at least at the macro level, of the Keynesian emphasis on wages being sticky downwards, with the implication that adjustment to shocks in the labour market comes less via changes in prices (wage rates) than via changes in quantities (employment and unemployment).

Even so, at the micro level, the syllabus carries an implicit acceptance of the neoclassical story that wage rates are set at the point where the marginal revenue product of labour curve crosses the labour supply curve and, of course, the market clears. A key explanatory variable is the elasticity of demand for labour, which is the effect on employment of a change in wages. This simple analysis is, of course, part of the model of perfect competition.

The unsuitability of the simple model

 But you don’t have to think hard before you realise that, when it comes to the labour market, the unsuitability of the simple neoclassical model runs a lot deeper than just the many respects in which all real-world markets fall well short of the assumptions of perfect competition. The most glaring respect in which the labour market differs from all other markets, whether markets for goods or markets for the factors of production, is that the rest involve the purchase or sale of inanimate objects, whereas each unit of labour purchased or sold comes with a human being inextricably attached. This obvious truth has many implications for the way labour markets work.

Perhaps the first person to formally note this truth was Alfred Marshall, the (British) father of neoclassical economics – the “marginalist revolution” – in one of the later editions of his magnum opus, The Principles of Economics, first published in 1890, which was the dominant textbook used in university economics courses throughout the English-speaking world until it was displaced by Keynes’s General Theory in the 1940s.

Surprisingly, given its history of neoclassical orthodoxy, this unique feature of labour – its inseparability from the humans delivering it – was readily acknowledged by the Productivity Commission in its report on the Workplace Relations Framework in late 2015. The report’s first “key point” is that “a workplace relations framework must recognise two enduring features of labour markets” the first of which is that “labour is not just an ordinary input. There are ethical and community norms about the way in which a country treats its employees” (page 2). This is true enough, but I’d have thought it was a strange way to put it. If you except that economies are run for the benefit of the people living in them, then the real point is not that it’s “unethical” to treat workers badly, but that the vast majority of people living in any economy are employees or their dependents. People are ends in themselves, not means to an end in the way that inanimate objects are merely means to human ends.

Further in (page 83), the PC says that Labour market outcomes do not just affect economic performance — they also have a substantial impact on equality of opportunity, the stability of family relationships and social cohesion more generally. The ethical and social dimensions of the labour market form the basis for many aspects of the WR system that differentiate it from the regulation of other markets.

“For example, the ‘price’ of labour differs from the price of most other inputs in an economy. A broad principle underpinning Australia’s competition policy framework is that lower prices from competition are almost always desirable. In labour markets it is less clear that a lower price is necessarily desirable, given that many people’s incomes and wellbeing depend to a considerable extent on the price of labour and it can be costly to use alternative mechanisms to redistribute income. Indeed, the existence of a minimum wage — a ‘floor price’ set by regulation, which would usually be seen as contrary to the public interest for other goods and services — illustrates this distinction.

In a nearby section of the report headed Human complexities (page 86), the PC acknowledges further implications of the labour market’s animate rather than inanimate nature:

In the real world, employers and employees are people with all their various flaws and virtues, and these can collide in workplaces in ways that have ramifications for how labour markets function:

• People make mistakes (for example, employers and employees may form an employment contract without any real due diligence).

• Employers and employees have values that are important to the way they do their work. An employee may want to work many additional hours at no cost because of professional pride. Employers may want to pay bonuses, provide better staff facilities or assist an employee facing family problems (say domestic violence) because they are dealing with human beings who they wish to help and please. Employers and employees dealing with each other are not merely doing so as part of a calculated business strategy, and in some cases this opens the door for one party to exploit the other’s goodwill and non-monetary motivations. (One less altruistic formulation of this is that employers may sometimes set higher prices for labour to motivate trust and to increase the cost of shirking — one example of so called ‘efficiency wages’.)

• There are few ‘representative’ employers and employees. People have heterogeneous tastes for workplace conditions and heterogeneous abilities, even when paid the same wage rate.

• Some businesses are poorly managed, and most are not at the technological and managerial frontier. An inadequately managed firm may provide poor training, treat people poorly, leave them bored or over busy with poor task scheduling, pay them too little for what they do, or provide no praise for good work — and yet people do not leave the first time they are ill-treated. On the other hand, there are model employers, with a spectrum of employers between the two extremes. The poorest performing employers may fail ultimately, but failure usually takes time, and damage in the interim may not be limited to just the employer. There is a persistent poorly performing tail in the distribution of firm performance in all countries and all industries.

Some of the above complexities suggest a need for regulation, others not. For example, regulation of unfair dismissal is justified, not only because the act itself is problematic but also because the potential to do it allows leverage by an employer to exploit vulnerable employees. Bullying would fall under the same category (whether by an employee or employer). Voluntary consent to work longer hours than the average is not an obvious problem, unless it is actually not ‘voluntary’, but obtained through coercion.”

Market failure in labour markets

The fact that the labour market is so personalised – workers are people, but so too are bosses – means that many of the usual respects in which all real-world markets differ from perfect competition, all the common or garden categories of market failure, are a lot more significant in the case of the labour market. For instance, why are wages sticky downwards as Keynes told us? Because of the human factor. Because employers know that cutting workers’ nominal wages makes them very unhappy and likely to be less enthusiastic about doing their jobs well. Much better during a recession to just leave nominal wages unincreased and wait for continuing inflation to reduce them in real terms. (This, by the way, is why sensible macro managers like to see some price inflation: it makes it easier to cut real wages when sometimes you need to.)

Speaking of the more standard reasons the labour market falls short of perfect competition, the PC’s report offers a good list of them (page 85):

• information asymmetries. Jobseekers may find it difficult to know the extent of competition for a job, the standard levels of remuneration and conditions for a comparable employment opportunity, and the non-wage conditions of a new workplace — such as workplace morale or the behaviour of managers. For employers, it may be similarly difficult to evaluate a potential employee’s skills or personal attributes, and other opportunities or offers the employee is considering. These gaps in information increase the uncertainty of rejecting an offer during negotiations. Even where parties can overcome these information gaps, this is likely to come at a significant cost.

• search costs. Job searching is costly, as is recruitment. It is also an uncertain process — parties usually make and receive offers in a sequential fashion, and so must consider the likelihood of receiving a better offer or applicant in the future. For workers whose skills or knowledge are not easily transferable between jobs, the financial costs and time taken to switch between employers or job sectors may be particularly high.

• impediments to individuals freely entering and exiting the labour market. Many people do not have sources of non-labour income or savings to support themselves if they do not work. Even where safety nets such as unemployment benefits are available (though they can be difficult to access), the personal and social costs of unemployment mean that many people may not see exiting unsatisfactory employment as a viable alternative.

• barriers that limit the mobility of labour between segments of the labour market. People can find it difficult to relocate to areas where jobs are more available, due to influences such as family circumstances, housing and ties to local communities and infrastructure. While developments such as long-distance commuting, temporary immigration, and advances in transport and communication technology have improved labour mobility in Australia, there are still significant personal reasons that hold employees to locations.

• employers that wield substantial purchasing power in the labour market (monopsonies). While monopsonies are historically associated with ‘one company towns’ where employees have little recourse to seek jobs nearby, they still persist in some sectors, for example government-provided services, or where the skills required by firms are sufficiently differentiated (sometimes referred to as monopsonistic competition). Behaviour to similar effect can also occur where employers in certain industries ‘cooperate’ to prevent wage bidding wars for talented employees.

These characteristics mean that in the absence of labour market regulations, wages are not necessarily set purely by reference to a competitive market rate, but rather through bilateral bargaining between employer and employee. The relative bargaining power of each party will determine their capacity to influence the final wage outcome.

Unequal bargaining power

Given the PC’s acknowledgement of the significance of all these departures from the neoclassical model, it’s not surprising that the second of the two “enduring features of labour markets” it highlights in the “key points” of the report is that without regulation and an ability to act collectively, many employees are likely to have much less bargaining power than employers, with adverse outcomes for their wages and conditions. Equally, poorly-designed regulation can risk bestowing too much power on organised labour in their dealings with individual employers. The challenge for a WR framework is to develop a coherent system that provides balanced bargaining power between the parties, that encourages employment, and that enhances economic efficiency. It is easy to both over and under regulate.

In its appendix H on bargaining power, the report says most [people] agree that the central goal of workplace relations policy is to reduce the superior bargaining power of employers over employees that would occur in the absence of regulation . . . 

• Bargaining power originates from the relative costs to contracting parties from failing to reach an agreement, with the result that one party can achieve leverage to re-distribute returns from the other.  For example, the cost of not employing a given employee may be low for the employer, while high for the employee.

• The neoclassical model of a perfectly competitive labour market predicts that imbalances in bargaining power cannot persist, with both employers and employees being ‘price takers’. However, there are a variety of factors that can differentiate the labour market (or at least many parts of it) from the perfectly competitive model, including: information asymmetries and search costs; a lack of voluntary entry and exit from the labour market; impediments to labour mobility; and employers with substantial purchasing power in the labour market (monopsonists).

• These factors mean that wages are generally set by employers and employees through bargaining, rather than purely by a competitive market rate. The resulting wage thus reflects relative differences in the bargaining power of parties.

• In the absence of regulation, imbalances in bargaining power would often be tilted towards employers, but in some circumstances may favour employees or unions.

• To counteract perceived inequalities in bargaining power, governments respond with policies such as enforcing minimum standards within the labour market (for example, minimum wages), and allowing employees to unionise and collectively bargain.

• A key regulatory concern is to ensure that in mitigating the risks of excessive employer bargaining power, regulations do not overcompensate by granting excessive power to employees.

Minimum wage

 Which brings us to the minimum wage. The PC’s position is that minimum wages are justified, and the view that existing levels are highly prejudicial to employment is not well founded. However, significant minimum wage increases pose a risk for employment, especially for more disadvantaged job seekers and in weakening labour markets.

The report goes on to note (page 177) that:

• Australia’s national minimum wage is high by international standards. It has risen in real terms over the last decade, although its growth rate has been constrained to reduce its ‘bite’ (the minimum wage as a share of median wages).

• There is an economic rationale for a regulated minimum wage that lifts the incomes of low-paid workers above the levels they would otherwise receive, to counter the effects of imbalances in bargaining power and other market distortions. There are also equity arguments . . . 

• At present, it is not possible to pinpoint the impacts of minimum wages on employment. Economic theory and some international empirical studies suggest that increases in minimum wages can reduce jobs and hours worked, but they also indicate that employment gains are possible in some circumstances. There have been few clear-cut wage ‘experiments’ in Australia and many studies are dated and/or have data and methodological limitations. The indirect evidence is also not clear-cut.

• While not definitive, the Productivity Commission’s assessment is that modest increases in Australia’s minimum wage are unlikely to measurably affect employment, but that large increases in minimum wages would reduce employment. How, and at what rate, such effects manifest will vary depending on economic conditions and other policy settings.

If you’re not surprised by all that, you should be. At least until the publication of an empirical study by Card and Krueger in 1993, most economists were sure a “binding” minimum wage – one that held the wage rate above the level market forces would have set – would cause employment to be lower. This is what the neoclassical model predicted, and there was little reason to doubt it was true. Today, however, economists are strongly divided on the question, with many now doubting that modest increases in the minimum wage do much if anything to affect employment, while clearly benefiting those already on the wage. The Card and Krueger study compared changes in employment levels in the fast-food industry in adjoining states, New Jersey and Pennsylvania, when one state increased minimum wages and the other didn’t. It found that, if anything, employment rose a fraction after wages were increased. So this age-old question has now become an empirical rather than a than a theoretical question. Many more empirical studies have been done since Card and Krueger, and while many have confirmed its broader conclusion that minimum wage increases have little effect on employment, many have found that there remains a quite small negative effect.

Alternative models of the labour market

When labour economists realised how unsuited the neoclassical model was to analysing the workings of the labour market – and how off-beam its predictions could be, they began developing alternative models, ones with more realistic assumptions and thus more credible predictions. Trouble is, while most of these alternatives offer more believable explanations of some aspects of the labour market, none is sufficiently comprehensive as to allow it to be adopted as a replacement to the simple neoclassical model with its (often wrong) answer for everything. This is a big part of the reason the old model remains influential in many areas in addition to analysis of the labour market. I’ll run you quickly through a few of the lesser models before coming to the two I think are most useful, the efficiency wage and the oligopsony model.

Drawing on RG Castles little text, the dual labour market theory developed by the American economist Michael Piore argues that labour markets can be split into two distinct sectors, primary and secondary labour markets. The primary market consists of stable, relatively well-paid jobs, usually requiring both qualifications and skills. The company invests in training and seeks a long-term relationship with the employee. The secondary sector, by contrast, is characterised by poorly paid jobs with minimal training and high levels of staff turnover.

A related idea is the concept of “internal labour markets”. Most workers in the primary labour market are sheltered from the effects of changes in supply and demand in the labour market. Once they obtain a primary job, their future depends on the operation of an internal labour market within the firm. Firms have a long-term investment in their employees through on-the-job training and wish to encourage workers to remain with the firm. Management encourages productivity within a framework of long-term job security.

Efficiency wage. This term was first used by Marshall in the last edition of his text in 1924, by was developed in a different direction by Carl Shapiro and Joe Stiglitz in 1984. It argues that, at least in some markets, wages form in a way that doesn’t clear the market. It points to the incentive for managers to pay their employees more than the market-clearing wage so as to increase their productivity or efficiency, or to reduce costs arising from staff turnover. This greater efficiency justifies the higher wage. Even so, if wage rates are above the market-clearing level, unemployment is persistent. Shapiro and Stiglitz developed the case where, in markets where it’s difficult to measure the quantity and quality of a worker’s effort, there is an incentive for managers to pay a higher wage to discourage “shirking”, since workers have more to lose if they were sacked for shirking. A different rational motive for paying higher wages occurs where the high cost of training replacement workers means paying a higher wage to discourage staff turnover is justified. Or, if workers abilities differ, paying higher wages should help the firm recruit and retain more-able workers. George Akerlof’s version argues that higher wages encourage high morale, which raises productivity.

Oligopsony model. Monopsony means one buyer of a product or, in this case, labour. Oligopsony means just a few buyers – by no means uncommon in a modern economy where a few big companies dominate many product markets. As explained by Alison Booth, the oligopsony model assumes that even if workers have identical skills and abilities, they have differing preferences on which employer they want to work for, influenced by such things as how far the firm is from where they live, the hours they want to work, or whether they like the boss and their fellow workers.

 It takes time and effort (that is, cost) for workers to find alternative employers they like at least as much as their present one and, similarly, it’s expensive for employers to find a worker they like as much as the one they could lose. This makes many workers reluctant to change jobs and many bosses reluctant to change workers. And because these preferences are private information – the other side can’t be sure how strong there are – there’s scope for “economic rents”: for workers to be paid less, or more, than the value of their work. Less is more likely. Booth says the attraction of the oligopsony model is its ability to show how a minimum wage can actually increase employment, as well as why employers provide general training to workers who could leave and take the training with them.